Upload
doanthuan
View
220
Download
0
Embed Size (px)
Citation preview
ENSC 833: NETWORK PROTOCOLS AND PERFORMANCE
FINAL PROJECT PRESENTATION
SPRING 2016
EFFICIENT HANDOVER IMPLEMENTATION IN LTE-BASED FEMTO-CELL ENVIRONMENT
Presenters: Instructor :
Nnamdi Eyisi ([email protected]) Prof. Ljiljana Trakojovic
Haris Mehmood ([email protected])
Website:
http://www.sfu.ca/~hmehmood4/25/2016
1
OVERVIEW
Introduction
Project overview
LTE
Femto Cells
Types of handover
Implementation of proposed algorithm
Handover Algorithm
Handover Parameters
Results
Discussion
References
4/25/2016
2
INTRODUCTION
4/25/2016
3
PROJECT OVERVIEW
Proposed algorithm is to be implemented in an LTE based femtocell environment
Femtocells have low cost and power requirement, and relatively small coverage area. Typically
private owned and deployed indoors
Ping-pong effect usually more frequent femtocells largely due to its small coverage area
Handover of interest is User Equipment's (UEs) moving between femtocells and macrocells
Proposed algorithm aims to reduce frequent ping-pong handover using occurring at the boundary of
the cell
User Equipment's (UEs) that are candidates for handover are categorised first based on their speed
UEs within a certain speed range are given priority for handover depending on the type of service
4/25/2016
4
PING-PONG EFFECT
Ping-pong is common phenomenon which occurs as result
of unnecessary handovers and can reduce the system
efficiency by as far as 40%.
Our aim is to eliminate this problem in high speed
Vehicular/medium speed traffic environment.
4/25/2016
5
WHAT IS LTE ?
LTE (Long-Term Evolution , commonly marketed as 4G LTE) is a standard for
wireless communication of high-speed data for mobile phones and data terminals.
It is based on the GSM/EDGE and UMTS/HSPA network technologies, increasing the capacity and
speed using a different radio interface together with core network improvements.
The standard is developed by the 3GPP (3rd Generation Partnership Project) and is specified in its
Release 8 document series, with minor enhancements described in Release 9.
4/25/2016
6
FEMTO-CELL OVERVIEW
Femto-cells are defined as the low power cells that are normally used for indoor radio
reception.
They are normally part of macro-cells.
5G technologies are currently in the process of using millimeter wave and would deploy
femto-cells in commercial environment.
Our decision to study femto-cells stem from the fact that they part an important part of next
generation networks.
4/25/2016
7
FEMTO CELL ARCHITECTURE
Figure retrieved from www.telecom-cloud.net 4/25/2016
8
HANDOVER IN LTE
X2 Handover
This type of handover occurs between two macro cells belonging to the same Core network. X-
2 interface exists between the two e-nodeB.
S1 Handover
This type of handover occurs between macro cells belonging to the different core networks. It
flows through the core network instead of the radio network.
S1 handover is implemented between femtocells and macrocells as communication
between macrocells (eNB) and femtocells (HeNB) are routed via the core network and
not through the X2-interface (direct communication)
4/25/2016
9
S-1 BASED HANDOVER
Figure retrieved from jwcn.eurasipjournals.springeropen.com
4/25/2016
10
IMPLEMENTATION
4/25/2016
11
LENA HANDOVER ALGORITHMS
RSRP Algorithm
Reference Signal Received Power (RSRP), is defined as the linear average over the power contributions (in [W]) of the resource elements that carry cell-specific reference signals within the considered measurement frequency bandwidth.For RSRP determination the cell-specific reference signals R0 according TS 36.211 [3] shall be used. If the UE can reliably detect that R1 is available it may use R1 in addition to R0 to determine RSRP.
RSRQ Algorithm
Reference Signal Received Quality (RSRQ) is defined as the ratio N×RSRP/(E-UTRA carrier RSSI), where N is the number of RB’s of the E-UTRA carrier RSSI measurement bandwidth. The measurements in the numerator and denominator shall be made over the same set of resource blocks.
Noop Algorithm
No handover algorithm.
4/25/2016
12
Base LENA configuration
Soft Frequency Reuse
In Soft Frequency Reuse (SFR) scheme each eNb transmits over the entire system bandwidth, but there are two sub-bands, within UEs are served with different power level. Since cell-center UEs share the bandwidth with neighboring cells, they usually transmit at lower power level than the cell-edge UEs. SFR is more bandwidth efficient than Strict FR, because it uses entire system bandwidth, but it also results in more interference to both cell interior and edge users.
Proportional Fair Mac Scheduler
The Proportional Fair (PF) scheduling supports high resource utilization while maintaining good fairness among network flows. A user is likely to be scheduled when its instantaneous channel quality is high relative to its own average channel condition over time
4/25/2016
13
FEMTO CELL CREATION IN NS-3
Macro-cell
• Create a standard Macro cell
Assign Building Block
• Use building size equal to femto cell area
Place Femto Cell
Transmitter
• Isotropic antenna, Building loss model, Fading model, low power
4/25/2016
14
RADIO ENVIRONMENT MAP
The radio environment map shows three distinct macro cells, each having three sectors.
The box shows the building covered by the femto cells.
4/25/2016
15
CELL PARAMETER SELECTION
PARAMETER
NAME MACRO CELL FEMTO CELL
Cellular layout 14 three-sectored sites in hexagonal layout(42 cells in total) Single Cell (4 Placed inside each Macro cell)
Inter-site distance 500m 50m
Cell Tx power 30 dBm 10 dBm
Path loss model L = 128:1 + 37:6 log10 R Channel fading Typical urban Indoor propagation model with fading
Carrier frequency 2 GHz 2 GHz
System bandwidth 5 MHz (25 RBs) 5 MHz (25 RBs)
Error model None None
UE distribution 14 UEs distributed randomly in front of each eNodeB (588 UEs in total) 6 in each femto cell
UE measurement interval 25 ms 25 ms
Simulation duration 1200 s 1200 s
4/25/2016
16
MOBILITY PARAMETER SELECTION
PARAMETER
NAME VEHICULAR PEDESTRIAN URBAN
UE Type Outdoor Indoor Outdoor Indoor Outdoor Indoor
UE average speed 38 km/h 5 km/h 10 km/h 4 km/h 24 km/h 4 km/h
UE direction Coverage Edge Random Indoor Random Edge Random Indoor Random Random
Direction Change 5s 3s 5s 3s Random Random
UE type percentage 80% 20% 60% 40% 45% 55%
UE distribution Cluster Random Cluster Random Random Random
UE movement pattern Random Indoor path Random Indoor path Random Random
UE Min. Speed 25 km/h 2 km/h 2 km/h 2 km/h 15 km/h 0
UE Max. Speed 60 km/h 7 km/h 20 km/h 7 km/h 60 km/h 6 km/h
4/25/2016
17
TRAFFIC MODELING
PARAMETER
NAME UDP TCP TCP
Application Type udp echo server Voice Constant bit rate Bulk Send Application
Real Time No Yes No
Description Ping server VoIP Send Max Bytes
QCI No 3 No
Start time 0.0 0.0 0.0
End time 1200 1200 1200
Interval No No Burst traffic Reset 60s
Bandwidth Available EPC Guaranteed bit rate Maximum available bit rate
4/25/2016
18
PSEUDO CODEInitialization
Check for handover trigger conditions.
If(Handover threshold satisfied)
{Track the mobile speed using the Doppler shift
If(UE Speed is determined)
{ if !(Cell Max capacity)
{
If (UE Speed < 15km/h)
{Perform handover
{else if (UE Speed <=30km/h)
Determine whether the application is real time
If(Real Application)
{ Perform handover}
else {no handover}}}
If(UE Speed > 30km/h)
{ no handover }
}
else { no handover}}} 4/25/2016
19
4/25/2016
20
SIMULATION SCENARIOS
Scenario 1:
Apply the algorithm to the vehicular environment.
Scenario 2:
Determine the effect of different handover parameters in the pedestrian environment.
Scenario 2(a):
Analyze the effect of Time to trigger on the handover performance.
Scenario 2(b):
Analyze the effect of Threshold hysteresis on the number of handover.
Scenario 2(c):
Analyze the effect of window interval on the handover performance.
Scenario 3:
Apply the optimized parameters and algorithm to the urban environment and analyze the results.
4/25/2016
21
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
Handover Comparison
No of Handovers per user per second in 60s time intervals No of Handovers per user per second using efficient Algorithm in 60s time intervals
The graph shows that per user handover decreases considerably for high speed users using our
implemented algorithm. The performance improvement is almost 40 percent.
1.VEHICULAR ENVIRONMENT
a). Handover Comparison using algorithm
4/25/2016
22
-110
-105
-100
-95
-90
-85
-80
RSRP trace of a High speed user
Cell1 macrocell Cell2 femtocell
0
0.05
0.1
0.15
0.2
0.25
0.3
RA
DIO
LIN
K F
AIL
UR
E %
SIMULATION INTERVAL
Radio Link Failures Percentage
The radio failure percentage is limited to 0.25 percent which shows a lot of users are still able to
receive good radio reception of the macro cells when they are crossing the boundary.
The second graph shows the SINR trace of a single high speed user. This shows that the user though
has degraded experience for a few seconds is able to continue normal service. Since, there is no
handover the user is continually scanning for radio signals from both eNodeb and Home eNodeb.
b). Radio link Statistics
4/25/2016
23
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
60 120 180 240 300 360 420 480 540 600 660 720 780 840 900 960 1020 1080 1140
Handover
200 ms 400 msTime to Trigger
Ha
nd
ov
erp
er u
ser
2.PEDESTRIAN ENVIRONMENT
a). Time to Trigger
The graph shows that per user handover decreases considerably for pedestrian traffic when the time to trigger is increased, as the eNodeb takes more time to initiate handover procedure once the UE increases the offset threshold
4/25/2016
24
b). Handover Hysteresis
0
0.1
0.2
0.3
0.4
0.5
0.6
Time Interval
60
120
180
240
300
360
420
480
540600
660
720
780
840
900
960
1020
1080
Handover per user at Multiple Hysteresis Values
Hys=1 db Hys=3 db Hys=6 db
Increasing the threshold hysteresis decreases the overall number of handovers. However, this may
results in more radio failures and performance degradation.4/25/2016
25
c). Window Size
0
0.1
0.2
0.3
0.4
0.5
0.6
Sliding Window Size
Sliding Window size = 200ms 400ms
The graph shows that sliding window has a positive effect in decreasing the number of ping pong handovers. The sliding window is the time period where eNodeb averages the user power measurements to make a decision on handover.
4/25/2016
26
RADIO LINK FAILURES FOR WORST
CASE
0
0.1
0.2
0.3
0.4
0.5
0.6
Radio Link Failures Percentage
4/25/2016
27
3.URBAN ENVIRONMENT
a). Handover Parameter Calculation
Urban radio network is a combination of vehicular and pedestrian environment.
Determine the value of optimized handover parameters on the basis of the preceding environment
behavior.
Hysteresis = 3db. The optimized value for choosing the handover threshold.
Window size = a x W(ped)+b x W(veh)
whereas W(ped) corresponds to sliding window value for pedestrian environment. W(veh) corresponds to
the value of the parameter in the vehicular environment.
Trigger time = a x T(ped)+b x T(veh).
whereas T(ped) corresponds to handover trigger timer value for pedestrian environment. T(veh)
corresponds to the value of the parameter in the vehicular environment.
a and b are the fraction of UEs’ in the pedestrian and vehicular environment.
We will choose W=300 ms and T= 300ms.
4/25/2016
28
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
Handover vs Radio Link Failures
Handover per user per interval Radio Link Failures %
b). Handover Performance evaluation
The graph shows that the radio link failures do not exceed our benchmark of 0.5%.4/25/2016
29
CONCLUSION
We have finished the detailed modeling using vehicular and pedestrian environment and combining them both to form urban environment. It has given us insights on the radio behavior.
We demonstrated that proposed handover algorithm is efficient in reducing unnecessary handovers
The proposed algorithm had radio link failure rate of less than 0.3%
The categorization of network users based on speed and type of traffic reduced the amount of resources reduces the number of handovers while having minimal impact on quality of service
Every urban environment can be modeled as the combination of the vehicular and the pedestrian environment.
The optimization of the handover can be done by following the step by step approach.
High Speed vehicles passing through the boundary of femto-macro cell has still enough radio coverage to continue service reception.
4/25/2016
30
ALGORITHM PROPOSAL
Let Tv and Tp be the trigger time for the vehicular and pedestrian environment.
Let Wv and Wp be the sliding window averaging for the vehicular and pedestrian environment.
Alpha and Beta represent the fraction of UEs’ moving with less and more than 30 km/h respectively.
‘e’ represents the error during the previous computation.
Ts and Ws represent the stable values of trigger time and window averaging respectively.
Ti and Wi represent the indicated values of the above mentioned parameters.
Pp is defined as the function of ping pong handovers and pedestrian traffic. All handover between same
Ues’ within 2 seconds are determined as ping pong.
Pv is defined as the function of ping pong handovers and vehicular traffic.
For every averaging interval,
Determine Ti=alpha*Tp+Beta*Tv-e*Ti Wi=alpha*Wp+Beta*Wv-e*Wi
Error ‘e’= (Pp*alpha+Pv*beta)/Total Handovers e is positive if more errors are caused by vehicular
traffic and negative otherwise.4/25/2016
31
FUTURE RECOMMENDATIONS
Neural network based algorithm can be developed for self optimizing network.
Doppler shift can be implemented for calculating the velocity on the e-nodeb level.
Femto to femto handover can be further explored which has been our of the scope of this project.
User-user communication is also an emerging handover algorithm which can be efficient.
Application layer traffic can be changed to focus the studies more on the impact of type of traffic and
handover.
4/25/2016
32
REFERENCES
4/25/2016
33
[1] U. Dampage and C. Wavegedara, "A low-latency and energy efficient forward handover scheme for LTE-
femtocell networks", 2013 8th IEEE International Conference on Industrial and Information Systems,
Peradeniya, 2013, pp.53-58.
[2] C. Lee and J. Kim, "System Information Acquisition Schemes for Fast Scanning of Femtocells in 3GPP LTE
Networks", IEEE Commun. Lett., vol. 17, no. 1, pp. 131-134, Jan. 2013.
[3] A. Rath and S. Panwar, "Fast handover in cellular networks with femtocells", 2012 IEEE International
Conference on Communications (ICC), Ottawa, ON, 2012, pp.131-134.
[4] I. Shayea, M. Ismail and R. Nordin, "Advanced handover techniques in LTE- Advanced system", 2012
International Conference on Computer and Communication Engineering , Kuala Lumpur, 2012, pp.74-79.
[5] M. Lee and S. K. Oh, "Fast handover scheme using handover notification with no acknowledgement," 2011
IEEE MTT-S International Microwave Workshop Series on Intelligent Radio for Future Personal Terminals,
2011.
[6] Y. Wang, W. Peng, X. Qui, “Automated Handover Optimization Mechanism for LTE Femtocells” 7th
International ICST Conference on Communications and Networking in China (CHINACOM), 2012
[7] C. Lin, K. Sandrasegaran, H. Ramli, R. Basukala, “Optimized Performance Evaluation Of LTE Hard
Handover Algorithm With Average RSRP Constraint” International Journal of Wireless & Mobile Networks
(IJWMN) Vol. 3, No. 2, April 2011.
4/25/2016
34
[8] LTE Network Architecture. Retrieved, from http://www.tutorialspoint.com/lte/lte_network_architecture.htm
[9] H. Lee and Y. Lin, “A cache scheme for femtocell reselection,” IEEE Commununication Letter, vol. 14, no. 1, pp. 27–29, 2010.
[10] J. Moon and D. Cho, “Efficient handoff algorithm for inbound mobility in hierarchical macro/femto cell networks,” IEEE Communication Letter, vol. 13, no. 10, pp. 755–757, 2009.
[11] E. Dahlman, "LTE 3G Long Term Evolution," Expert Radio Access Technologies Ericsson Research, March 27, 2007.
[12] Z. Naizheng and J. Wigard„" On the Performance of Integrator Handover Algorithm in LTE Networks," in Proc. IEEE Vehicular Technology Conference. (VTC), pp. 1-5, 2008.
[13] I. Shayea et al,"Advanced handover techniques in LTE- Advanced system," in Proc. International Conference. on Computer and Communication. Engineering, pp.74-79, 2012.
[14] A. Rath et al., "Fast handover in cellular networks with femtocells," in Proc. 2012 IEEE International. Conference on Commununication. (ICC), pp. 2752-2757, June 2012.
[15] M.S. Shbat and V.Tuzlukov, "Handover technique between femtocells in LTE network using collaborative approach," in Proc. Asia-Pacific Conf. on Commun. (APCC), pp. 61-66, 2012.
[16] V. Chandrasekhar, "Femtocell networks: a survey," IEEE Commun.Mag., vol. 46, pp. 59-67, Sept 2008.
[17] C. Jiming, et al," Adaptive soft frequency reuse scheme for inbuilding dense femtocell networks," China Commun., vol. 10, no.1, pp. 44-55, 2013.
[18] K. Ivanov, G. Spring, ‘Mobile speed sensitive handover in a mixed cell environment’. Vehicular Technology Conference, July 1995, vol. 2, pp. 892–896
[19] W. Luo, X. Fang, M. Cheng, X. Zhou, ‘‘an optimized handover trigger scheme in LTE systems for high-speed railway’’. Signal Design and its Applications in Communications (IWSDA), October 2011, pp. 193–196
[20] P. Legg, G. Hui, J. Johansson, “A simulation study of LTE intra-frequency handover performance”. Vehicular Technology Conference Fall (VTC), September 2010, pp. 1–5.
[21] LENA ns-3 Retrieved, from http://www.nsnam.com
4/25/2016
35
QUESTIONS ?
4/25/2016
36